BenchLeader
Menu
LeaderboardsModelsProvidersBenchmarksCompareReleases1BlogMethodology

Vals Multimodal Index

Vals AI's GDP-weighted multimodal composite. Run by Vals AI.

As of 19 Sept 2026, Claude Fable 5 leads Vals Multimodal Index on BenchLeader with 74.2%, ahead of Claude Opus 5 at 73.9%, across 31 model configurations with a published result.

Published by
Vals AI
Category
Composite
Index weight
Reference only
Models
31
Data as of
19 Sept 2026

Vals AI (vals.ai).

What the test looks like

A composite of Vals AI's image-based benchmarks, weighted by the economic value of the underlying task types.

How it is scored

Vals AI's composite index.

What to keep in mind

A composite; read the component boards for detail.

Top 15
  1. 1Claude Fable 574.2%
  2. 2Claude Opus 573.9%
  3. 3Kimi K373.4%
  4. 4GPT-5.6 Sol (max)72.6%
  5. 5Claude Opus 4.870.9%
  6. 6Muse Spark 1.269.8%
  7. 7GPT-5.6 Luna (max)69.1%
  8. 8Claude Sonnet 568.8%
  9. 9GPT-5.5 (xhigh)68.1%
  10. 10Claude Opus 4.767.4%
  11. 11Muse Spark 1.1 (xhigh)66.7%
  12. 12Qwen3 8 (max)65.4%
  13. 13Gemini 3.6 Flash (high)65.1%
  14. 14GPT-5.6 Terra (xhigh)65.1%
  15. 15Grok 4.5 (high)63.4%
31 of 31
#
1Claude Fable 5Anthropic74.2%68.32026-08-11
2Claude Opus 5Anthropic73.9%66.72026-08-11
3Kimi K3Moonshot AIopen ↗73.4%64.22026-08-11
4GPT-5.6 SolmaxOpenAI72.6%68.82026-08-11
5Claude Opus 4.8Anthropic70.9%61.92026-08-11
6Muse Spark 1.2Meta69.8%–2026-08-11
7GPT-5.6 LunamaxOpenAI69.1%60.22026-08-11
8Claude Sonnet 5Anthropic68.8%57.22026-08-11
9GPT-5.5xhighOpenAI68.1%67.62026-08-11
10Claude Opus 4.7Anthropic67.4%64.52026-08-11
11Muse Spark 1.1xhighMeta66.7%61.82026-08-11
12Qwen3 8maxAlibaba65.4%66.52026-08-11
13Gemini 3.6 FlashhighGoogle65.1%61.72026-08-11
14GPT-5.6 TerraxhighOpenAI65.1%64.32026-08-11
15Grok 4.5highxAI63.4%61.02026-08-11
16Gemini 3.5 FlashhighGoogle62.9%63.62026-08-11
17Claude Sonnet 4.6Anthropic60.6%59.02026-08-11
18MiniMax-M3MiniMaxopen ↗60.0%57.52026-08-11
19Kimi K2.6Moonshot AIopen ↗56.4%60.52026-08-11
20Gemini 3.1 ProhighGoogle56.1%60.72026-08-11
21GPT-5.4 minixhighOpenAI54.2%55.72026-08-11
22Qwen3.7 PlusAlibaba53.9%59.92026-08-11
23Gemini 3 FlashhighGoogle52.2%58.72026-08-11
24Qwen3.6 PlusAlibabaopen51.5%58.52026-08-11
25Inkling SmallThinking Machinesopen ↗50.0%56.32026-08-11
26GPT-5.4 nanohighOpenAI47.6%49.12026-08-11
27Grok 4.3highxAI43.3%58.22026-08-11
28Claude Haiku 4.5thinkingAnthropic42.9%47.52026-08-11
29Gemini 3.1 Flash LitehighGoogle41.4%49.12026-08-11
30Grok 4.20thinkingxAI39.1%59.92026-08-11
31Mistral Medium 3.5highMistral AIopen34.8%–2026-08-11

Cite as: BenchLeader, “Vals Multimodal Index leaderboard”, https://www.benchleader.com/benchmarks/vals_multimodal_index, data as of 19 Sept 2026.

Vals Multimodal Index: questions

What does Vals Multimodal Index measure?
A composite of Vals AI's image-based benchmarks, weighted by the economic value of the underlying task types. Scores are reported in percent of tasks solved; higher is better.
Which AI model leads Vals Multimodal Index?
Claude Fable 5 leads Vals Multimodal Index with 74.2% as of 19 Sept 2026, ahead of Claude Opus 5 at 73.9%.
How many models have Vals Multimodal Index results?
31 model configurations have a Vals Multimodal Index result on BenchLeader, all taken from Vals AI.
Who runs Vals Multimodal Index and how often is it updated?
Vals Multimodal Index is published by Vals AI. BenchLeader re-reads the published results every morning and records the date each result was published.
Does Vals Multimodal Index count toward the BenchLeader Index?
No. Vals Multimodal Index is shown for reference but left out of the composite index.